Premium
Efficiency of an Incomplete Block Design Based on Geostatistics for Tillage Experiments
Author(s) -
López María V.,
Arrúe José L.
Publication year - 1995
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1995.03615995005900040023x
Subject(s) - tillage , geostatistics , block design , block (permutation group theory) , mathematics , plough , silt , mean squared error , statistics , soil science , sampling (signal processing) , spatial variability , environmental science , computer science , geology , agronomy , geomorphology , geometry , combinatorics , biology , filter (signal processing) , computer vision
Spatial dependence of soil properties often reduces the power of conventional statistical methods to detect treatment differences. Control of adverse effects of soil variability is of special interest in long‐term experiments when small and slowly developing treatment effects are generally expected to occur, as in tillage research. The main objective of this study was to evaluate the ability of an incomplete block design based on geostatistical concepts to improve the precision of a conservation tillage experiment conducted at four sites in Aragón, northeast Spain. A preliminary geostatistical characterization of plow layer variability at these sites showed that, in most instances, soil water content and silt plus clay content were spatially dependent. Maps of kriged estimates of these properties were used to locate the tillage plots according to the proposed design. Using incomplete blocks of Size 2, the method estimates treatment effects by making short‐distance comparisons and ensures spatially balanced treatment contrasts through fixed comparison distances. This design was compared with a complete block design using soil and crop data obtained during the first two growing seasons of the tillage experiment. The results of a total of 1050 analysis of variance comparisons revealed that the incomplete block design was, on average, 24% more efficient than the complete block design. The use of incomplete blocks reduced the average error mean square by 33% and increased the number of cases with significant tillage treatment differences by 25% compared with the use of complete blocks.